2,597 reputation
1321
bio website
location
age
visits member for 2 years, 3 months
seen 5 hours ago

I'm the Rose part of:

Rose, C. and Smith, M.D. (2002-2013)
Mathematical Statistics with Mathematica
Springer-Verlag / mathStatica

and one of the original developers of the mathStatica software add-on for Mathematica. I'm a long-time Mma fan since v2, ... a past Visiting Scholar at Wolfram Research, ... a recent guest editor of The Mathematica Journal, ... and a current editor of the Journal of Statistical Software.


Jul
19
answered Transforming a CopulaDistribution only works with the Product Copula
Jul
15
reviewed Approve suggested edit on Simulated Annealing for NMinimize[] always returning the same solution
Jul
11
comment Mma 10: Half the parallel power (Mac Pro)?
* v9 SINGLE-core: NPKDEPlot[data, bandwidths, {15, 25}] // AbsoluteTiming -----------> 23 seconds $$\text{ }$$ * v10 SINGLE-core: NPKDEPlot[data, bandwidths, {15, 25}] // AbsoluteTiming -----------> 42 seconds $$\text{ }$$As a guess, the v9 vs v10 timing difference might be due to increased plot points used under v10 (just a guess??) ... but the v9 vs v10 timing is not the issue here: it is the v10 timings on the same computer that are the subject of interest.
Jul
11
awarded  Nice Question
Jul
11
comment Mma 10: Half the parallel power (Mac Pro)?
@rcollyer mathStatica's NPKDEPlot function uses ParallelTable[Plot[ Funky ] ] to produce each separate curve on a separate kernel, where Funky is a Compiled application of (Map of Total of some Ifs and Buts).
Jul
11
comment Mma 10: Half the parallel power (Mac Pro)?
Can't agree. It is not the role of mathStatica, or any other package, to manually override the number of cores or kernels on an end-user's machine. That is Mathematica's role to set optimally ... it worked correctly in v9, and it worked correctly in pre-release versions of v10 .. and now it is suboptimal. Moreover, I can't agree with your assessment that it makes no difference, or that shutting down half your kernels is 'beneficial', because it plainly is not. Having multiple kernels for massively parallel problem is entirely the point of parallelisation. And we have gone backwards here.
Jul
11
revised Mma 10: Half the parallel power (Mac Pro)?
added 20 characters in body
Jul
11
comment Mma 10: Half the parallel power (Mac Pro)?
Have added real world test -> please see above.
Jul
11
revised Mma 10: Half the parallel power (Mac Pro)?
added real world test
Jul
11
asked Mma 10: Half the parallel power (Mac Pro)?
Jul
10
comment How to get the new font in Mathematica 10?
Ugly. I am almost tempted to use the term fugly, but shall refrain from doing that.
Jul
10
comment Fitting data to a Poisson distribution
If your data is discrete, then there is a MASS at each point $x = 0, 1, 2, 3$, etc ... and nothing inbetween. Accordingly, you should not be using a Histogram that places mass over the say 1.5 to 2.5, as if your data is continuous ... because it is not.
Jul
5
reviewed Approve suggested edit on Debuging in Mathematica
Jul
5
reviewed Reject suggested edit on Deleting all characters in a string that do not belong to a small set of elements
Jul
5
reviewed Approve suggested edit on Saner alternative to ContourPlot fill
Jul
5
reviewed Approve suggested edit on Problem printing what appears in my notebook
Jul
4
reviewed Approve suggested edit on How can I export specific data into different pages in .PDF format?
Jul
4
reviewed Approve suggested edit on How to export large matrix from .nb file into .pdf
Jul
2
awarded  Curious
Jul
1
reviewed Approve suggested edit on Drawing pairs (graph, circle packing)